Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Deep ViT Features as Dense Visual Descriptors

About

We study the use of deep features extracted from a pretrained Vision Transformer (ViT) as dense visual descriptors. We observe and empirically demonstrate that such features, when extractedfrom a self-supervised ViT model (DINO-ViT), exhibit several striking properties, including: (i) the features encode powerful, well-localized semantic information, at high spatial granularity, such as object parts; (ii) the encoded semantic information is shared across related, yet different object categories, and (iii) positional bias changes gradually throughout the layers. These properties allow us to design simple methods for a variety of applications, including co-segmentation, part co-segmentation and semantic correspondences. To distill the power of ViT features from convoluted design choices, we restrict ourselves to lightweight zero-shot methodologies (e.g., binning and clustering) applied directly to the features. Since our methods require no additional training nor data, they are readily applicable across a variety of domains. We show by extensive qualitative and quantitative evaluation that our simple methodologies achieve competitive results with recent state-of-the-art supervised methods, and outperform previous unsupervised methods by a large margin. Code is available in dino-vit-features.github.io.

Shir Amir, Yossi Gandelsman, Shai Bagon, Tali Dekel• 2021

Related benchmarks

TaskDatasetResultRank
Semantic CorrespondenceSPair-71k (test)
PCK@0.133.3
122
Semantic CorrespondencePF-PASCAL
PCK @ alpha=0.162.4
98
Video Object SegmentationDAVIS
J Mean52.1
58
Unsupervised Object DiscoveryCOCO 20k
CorLoc57.99
56
Unsupervised Object DiscoveryPASCAL VOC 2012
CorLoc71.64
28
Unsupervised Object DiscoveryPASCAL VOC 2007
CorLoc68.27
28
Video Object SegmentationDAVIS (val)
Mean J & F Score50.9
28
Novel View SynthesisD-RE10K static regions only (test)
PSNR18.67
26
Novel View SynthesisD-RE10K-iPhone full-image fidelity (test)
PSNR17.96
26
Semantic MatchingTSS
PCK (FG)64.7
24
Showing 10 of 17 rows

Other info

Follow for update